Those who support the idea of driverless cars say the computers that run them can avoid the mistakes that people make.

支持無人駕駛車輛概念的人們表示，操控這些車輛的電腦能夠避免這些人們會犯下的錯誤。

But those have problems as well.

但它們也有自己的問題。

Besides the potential for failures, malfunctions, errors, camera problems in heavy rain, one major concern of critics is: What if the computers of driverless cars are hacked?

除了可能出現的故障、失靈、錯誤、大雨下的相機問題，批評者主要關注的其中一個重點是： 如果無人駕駛汽車的電腦被駭客攻擊怎麼辦？

There may not be an answer for that yet, but the driverless technology itself continues to evolve.

也許我們現在對此還沒有解答，但無人駕駛技術本身仍在不斷發展。

Autonomous driving, we're told, is the way of the future.

我們常聽說自動駕駛是未來大勢所趨。

Autonomous cars featured heavily at this year's Paris Motor Show, but how can a computer deal with the randomness, the chaos of Parisian traffic?

自動駕駛汽車在今年的巴黎車展上大放異彩，但電腦要如何應對巴黎那隨機而混亂的交通呢？

We decided to put that to the test.

我們決定對此進行測試。

Our co-pilots today are engineers developing the technology that's already being sold to car makers. And where better to start than at the foot of the Eiffel Tower?

與我們一同坐在車內的是已經發展出這項技術，並將其賣給汽車製造上的工程師們。而沒什麼比從艾菲爾鐵塔底下出發更適合今天的主題了。

So now cars driving itself? - Yes.

所以現在車是自己在開？ - 對。

You're not involved at all? - No.

你完全沒有操控？ - 沒有。

No pedal, no wheel, no nothing? - No, no, no.

沒有踩油門或煞車、沒有控制方向盤，什麼都沒有？ - 完全沒有。

I got my hands on my knees.

我的手還放在膝蓋上呢。

Which should be a dangerous thing, given the cyclists, scooters, and pedestrians, not to mention the bad drivers who do have their hands on the wheel.

而這本來是相當危險的行為，畢竟周圍有著腳踏車、機車還有行人，更別提手雖然放在方向盤上但駕駛技術拙劣的汽車駕駛們。

But Beauvau says the car sensors are more efficient than the human brain.

但 Beauvau 說汽車上的感測器比人腦還更有效率。

And as you can see, there is a pedestrian here, and we're gonna slow because it's detected.

如你所見，前方有一個行人，而車子偵測到他所以慢了下來。

And then when the pedestrian is off the crossing area, then —

等到行人走過穿越區域之後就——

And even if the pedestrian come much faster, the car would know that. It's programmed to stop. - Yes.

而且就算行人用更快的速度衝出來，車子也會知道。它內部的程式就是設計成會讓它停下來。 - 沒錯。

The technology is part-programmed, part-learned.

這種科技部分預先以程式規劃，部分則是後天學習。

Through its many cameras, sensors, computers, and radars, the car's artificial intelligence allows it to learn as it goes.

藉由車上許多的鏡頭、感測器、電腦與雷達，車輛的人工智慧能夠在行進間進行學習。

It's just always focus on its own task, which is driving me safely from point A to point B.

它只專注在自己的任務上，也就是將我安全的從 A 點載到 B 點。

So that car did something very rude. - Yeah.

剛才那輛車做了超沒禮貌的動作。 - 是啊。

It just cut across you and the car felt it. - Yes.

它就直接這樣插過來，而這台車其實有感覺到。 - 對。

And it didn't even complain. - No.

而它甚至沒有抱怨。 - 沒有。

There was no tooting of villain. - No, ha ha ha.

也沒有憤怒地按幾下喇叭。 - 沒有，哈哈哈。

There will always be an element of risk going there, even for computers.

就算是電腦，也依然會有風險因素存在。

The zero fault doesn't exist.

完全沒有錯誤是不可能的。

What we want to show is that we're able to drive during one billion hours without any problem.

我們想要展示的是，我們能夠毫無問題地持續駕駛十億小時。

And that's much better than human beings. - Yeah, of course.

而這比人類的表現還要好上許多。 - 是的，當然。

Perhaps the most surprising thing about all this is that this is likely to be an evolution rather than a revolution.

也許這一切最令人驚訝的是，這很可能只是一場進化，而不是一場革命。

Already, all of the sensors that exist on this car and then allow it to drive autonomously exist on the sorts of cars that you would buy today. Ordinary cars.

在這款汽車上讓它能夠自動駕駛的所有感測器，都能在今天你能購買到的各類汽車上找到。普通的汽車。

And so what's likely to happen is that little by little, we will get in the habit of letting go of the steering wheel until one day, all cars, even here in the French capital, drive themselves.

所以可能會發生的是，我們將會一點一點的慢慢放開方向盤，而到有一天，即使在法國首都這裡，所有的車輛也都會是自動駕駛的。

Melissa Bell, CNN, Paris.

Melissa Bell，CNN 在巴黎報導。